Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A method of characterizing a specimen contained within a specimen container, comprising: providing classified 2D data sets obtained by processing a plurality of 2D images of the specimen container containing a specimen taken from multiple viewpoints, the classified 2D data sets being classified as: serum or plasma, settled blood portion, gel separator, air, tube, and label; correlating locations in the classified 2D data sets to a consolidated 3D data set; and forming a consolidated 3D model based upon the consolidated 3D data set.
This invention relates to automated characterization of biological specimens within containers, such as blood collection tubes, using 2D imaging and 3D modeling. The problem addressed is the need for accurate, non-invasive identification and segmentation of different components within a specimen container, such as serum, plasma, settled blood, gel separators, air, and container materials like tubes and labels. The method involves capturing multiple 2D images of a specimen container from different viewpoints. These images are processed to generate classified 2D data sets, where each pixel or region is labeled as one of the following: serum or plasma, settled blood portion, gel separator, air, tube material, or label. The classified 2D data sets are then correlated to a consolidated 3D data set, which integrates spatial information from the multiple viewpoints. Finally, a consolidated 3D model of the container's contents is formed, providing a detailed representation of the specimen's composition and distribution within the container. This approach enables automated analysis of specimen integrity, separation quality, and other characteristics without physical intervention. The method is particularly useful in clinical and laboratory settings for quality control and sample preparation.
2. The method of claim 1 , wherein the plurality of 2D images are taken at multiple different exposure times at each of the multiple viewpoints.
This invention relates to a method for capturing and processing 2D images from multiple viewpoints to enhance imaging quality, particularly in challenging lighting conditions. The method addresses the problem of limited dynamic range in imaging systems, where bright and dark regions in a scene cannot be captured simultaneously with sufficient detail. The method involves capturing a plurality of 2D images from multiple viewpoints, with each viewpoint capturing images at multiple different exposure times. By varying the exposure times, the system can capture both bright and dark regions of the scene with appropriate detail. The images from different exposures are then combined to produce a high dynamic range (HDR) image at each viewpoint. This approach allows for better visualization of scenes with high contrast, such as outdoor environments with bright sunlight and deep shadows. The method further includes processing the captured images to align and merge them, ensuring that the final output retains high fidelity and minimizes artifacts. The use of multiple viewpoints enables the creation of a 3D representation or a depth map, which can be used for applications such as augmented reality, robotics, or 3D reconstruction. The combination of multi-exposure imaging and multi-viewpoint capture enhances the overall imaging performance, providing a more comprehensive and detailed representation of the scene.
3. The method of claim 2 , wherein the multiple different exposure times comprise between about 0.1 ms and about 256 ms.
This invention relates to imaging systems that capture images using multiple different exposure times to improve dynamic range or other image quality metrics. The method involves acquiring a sequence of images of a scene with varying exposure durations, where each exposure time in the sequence falls within a range of approximately 0.1 milliseconds to 256 milliseconds. The exposure times are selected to optimize the capture of details in both bright and dark regions of the scene, allowing for subsequent processing to combine the images into a final output with enhanced dynamic range or reduced noise. The method may be applied in digital cameras, surveillance systems, or other imaging applications where capturing high-quality images under varying lighting conditions is important. The exposure time range ensures flexibility in adapting to different lighting scenarios, from very bright to very dim environments, while maintaining image clarity and detail. The technique may also be used in conjunction with other image processing steps, such as alignment, fusion, or tone mapping, to further enhance the final image quality.
4. The method of claim 2 , wherein the plurality of 2D images are taken at multiple different spectra having different nominal wavelengths.
This invention relates to imaging systems that capture multiple two-dimensional (2D) images of a target object at different spectral wavelengths. The method involves acquiring a set of 2D images, where each image is captured at a distinct nominal wavelength. By analyzing these images across different spectra, the system can extract detailed spectral information about the object, enabling applications such as material identification, defect detection, or environmental monitoring. The use of multiple wavelengths allows for enhanced contrast and resolution compared to single-wavelength imaging, improving the accuracy of spectral analysis. The system may include a tunable light source or multiple light sources with different wavelengths, along with a detector capable of capturing images at each selected wavelength. The captured images are processed to generate a spectral profile or map of the object, which can be used for further analysis or decision-making. This approach is particularly useful in fields like remote sensing, medical imaging, and industrial inspection, where spectral information provides valuable insights into the composition or condition of the target.
5. The method of claim 4 , wherein the multiple different spectra comprise three or more wavelengths between about 400 nm and 700 nm.
This invention relates to a method for analyzing biological samples using optical spectroscopy, specifically in the visible light range. The method addresses the challenge of accurately detecting and quantifying biological materials by leveraging multiple distinct wavelengths of light to improve measurement precision and reliability. The technique involves illuminating a biological sample with light at three or more different wavelengths within the visible spectrum, ranging from approximately 400 nm to 700 nm. This multi-wavelength approach enhances the ability to distinguish between different biological components and reduces interference from background noise or other substances. The method may be used in applications such as medical diagnostics, environmental monitoring, or food safety testing, where precise identification and quantification of biological materials are critical. By utilizing multiple wavelengths, the system can compensate for variations in sample composition and environmental conditions, leading to more accurate and consistent results. The technique may be integrated into portable or laboratory-based spectroscopic devices for real-time analysis.
6. The method of claim 4 , wherein the multiple different spectra comprise wavelengths of about 634 nm+/−35 nm, about 537 nm+/−35 nm, and about 455 nm+/−35 nm.
This invention relates to a method for analyzing biological samples using multiple different spectra to enhance detection and characterization. The method addresses the challenge of accurately identifying and quantifying biological materials, such as cells or molecules, by leveraging specific wavelength ranges to improve sensitivity and specificity. The technique involves illuminating a sample with light at three distinct wavelength bands: approximately 634 nm (±35 nm), 537 nm (±35 nm), and 455 nm (±35 nm). These wavelengths are selected to target different biological components, such as proteins, nucleic acids, or cellular structures, allowing for comprehensive analysis. The method may include preprocessing the sample, capturing spectral data from the illuminated sample, and processing the data to extract meaningful information. The use of multiple spectra enhances the ability to distinguish between different types of biological materials and improves the accuracy of measurements. This approach is particularly useful in applications like medical diagnostics, environmental monitoring, or biological research, where precise identification and quantification of biological samples are critical. The method may be integrated into existing analytical systems or used as part of a broader diagnostic workflow.
7. The method of claim 4 , wherein the classified 2D data sets are derived from optimally-exposed image data for each wavelength at multiple different exposure times.
This invention relates to image processing techniques for enhancing the quality of multi-wavelength imaging systems, particularly in scenarios where varying exposure times are used to capture image data. The core problem addressed is the challenge of obtaining high-quality 2D data sets from images captured at different wavelengths and exposure times, where optimal exposure settings may differ for each wavelength. The invention improves upon prior methods by ensuring that the classified 2D data sets are derived from optimally-exposed image data for each wavelength, captured at multiple distinct exposure times. This approach allows for better handling of dynamic range and noise across different wavelengths, leading to more accurate and reliable image classification and analysis. The method involves capturing image data at multiple exposure times for each wavelength, selecting the optimally-exposed data for each wavelength, and then classifying the resulting 2D data sets. This ensures that the final classified data retains the highest possible quality for each wavelength, improving the overall performance of the imaging system in applications such as medical imaging, remote sensing, or scientific research. The technique is particularly useful in systems where different wavelengths require different exposure settings to avoid saturation or underexposure, thereby enhancing the accuracy of subsequent data analysis.
8. The method of claim 1 , wherein the plurality of 2D images represent a 360 degree view of the specimen container that is based on multiple lateral images, with each lateral image overlapping adjacent images.
This invention relates to a method for capturing and processing 360-degree views of a specimen container using overlapping lateral images. The method addresses the challenge of obtaining a comprehensive, high-resolution view of a container from multiple angles without gaps or misalignment. The system captures a series of 2D lateral images of the container, where each image overlaps with adjacent images to ensure continuous coverage. These overlapping images are then stitched together to form a seamless 360-degree representation of the container. The overlapping regions allow for precise alignment and correction of distortions, ensuring accurate reconstruction of the container's surface. This method is particularly useful in applications requiring detailed inspection, such as medical or industrial container analysis, where a full circumferential view is necessary for quality control or diagnostic purposes. The overlapping images provide redundancy, improving the reliability of the final 360-degree view by allowing for error correction during the stitching process. The technique may also include preprocessing steps to enhance image quality, such as noise reduction or contrast adjustment, before stitching. The resulting 360-degree view can be used for automated analysis, such as detecting defects, measuring dimensions, or identifying surface features. The method ensures that no part of the container is obscured or misaligned, providing a complete and accurate representation for further analysis.
9. The method of claim 1 , wherein a number of the multiple viewpoints comprises 3 or more.
A system and method for capturing and processing images from multiple viewpoints to generate a three-dimensional (3D) representation of an object or scene. The technology addresses the challenge of accurately reconstructing 3D models from two-dimensional (2D) images by using multiple viewpoints to enhance depth perception and spatial accuracy. The method involves capturing images from at least three distinct viewpoints, where each viewpoint provides a unique perspective of the object or scene. These images are then processed to extract depth information, align the viewpoints, and generate a coherent 3D model. The use of three or more viewpoints improves the robustness of the reconstruction by reducing ambiguities in depth estimation and enhancing the accuracy of the final 3D representation. The system may include image sensors, processing units, and software algorithms for viewpoint alignment, depth mapping, and model generation. This approach is applicable in fields such as computer vision, augmented reality, medical imaging, and industrial inspection, where precise 3D reconstructions are required. The method ensures that the generated 3D model is detailed and accurate, enabling applications that rely on spatial data for analysis, visualization, or interaction.
10. The method of claim 1 , comprising computing statistics of optimally-exposed pixels at different wavelengths to generate statistical data.
A method for analyzing image data involves computing statistical properties of optimally-exposed pixels across different wavelengths to generate statistical data. The process begins by capturing an image using a sensor array, where each pixel in the array is sensitive to a specific wavelength of light. The sensor array includes multiple pixels, each configured to detect a distinct wavelength, allowing for the capture of multi-spectral or hyperspectral image data. The method then identifies optimally-exposed pixels within the captured image, which are pixels that have been exposed to light in a manner that provides the most accurate or useful data for analysis. These optimally-exposed pixels are selected based on predefined criteria, such as exposure time, signal-to-noise ratio, or other image quality metrics. Once identified, the method computes statistical properties of these optimally-exposed pixels, such as mean, variance, or other statistical measures, for each wavelength. The resulting statistical data provides insights into the distribution and characteristics of the image data across different wavelengths, which can be used for applications such as material identification, environmental monitoring, or medical imaging. The method ensures that the statistical analysis is based on the most reliable pixel data, improving the accuracy and robustness of the results.
11. The method of claim 10 , wherein the computing statistics of the optimally-exposed pixels from optimally-exposed image data for the different wavelengths comprises calculating a mean value, a standard deviation, and/or covariance from a collection of corresponding pixels from each wavelength.
This invention relates to image processing, specifically to analyzing optimally-exposed image data to compute statistical metrics for different wavelengths. The problem addressed is the need for accurate statistical analysis of image data across multiple wavelengths to improve image quality or enable further processing. The method involves processing image data that has been captured under optimal exposure conditions for different wavelengths. For each wavelength, corresponding pixels from the optimally-exposed image data are collected. From this collection, statistical metrics such as mean value, standard deviation, and covariance are calculated. These metrics provide insights into the distribution and relationships of pixel values across the wavelengths, which can be used for applications like color correction, noise reduction, or image enhancement. The method ensures that the statistical analysis is performed on high-quality, optimally-exposed data, leading to more reliable results. By computing multiple statistical measures, it enables a comprehensive understanding of the image data's characteristics across different wavelengths. This approach is particularly useful in fields like medical imaging, remote sensing, or scientific imaging, where accurate wavelength-specific analysis is critical. The calculated statistics can be used to improve image reconstruction, calibration, or other downstream processing tasks.
12. The method of claim 10 , wherein selection of the optimally-exposed pixels comprises selection of pixels from the images that include intensities of between about 180-254 based upon a range of 0-255.
This invention relates to image processing, specifically optimizing image exposure by selecting optimally-exposed pixels from multiple images. The problem addressed is the difficulty in capturing high-quality images under varying lighting conditions, where some regions may be overexposed or underexposed. The solution involves capturing multiple images with different exposure settings and then selecting pixels from these images that fall within an optimal intensity range to construct a final, well-exposed image. The method includes capturing a sequence of images with varying exposure levels. From these images, pixels are selected based on their intensity values, specifically those within a predefined range of 180-254 on a 0-255 scale. This range is chosen to exclude overly dark or overly bright pixels, ensuring the selected pixels are well-exposed. The selected pixels are then combined to form a final image with improved exposure quality. The process may involve additional steps such as aligning the images to correct for motion or distortion and applying blending techniques to ensure smooth transitions between selected pixels. The invention aims to enhance image quality by leveraging multiple exposures to capture the best possible pixel data, particularly in challenging lighting conditions. This approach is useful in photography, surveillance, and other applications where consistent exposure is critical.
13. The method of claim 10 , wherein a multi-class classifier is used to generate the classified 2D data sets.
A method for classifying 2D data sets using a multi-class classifier to improve accuracy and efficiency in data analysis. The technique addresses the challenge of accurately categorizing complex 2D data, such as images or sensor outputs, where traditional binary classifiers may fail to distinguish between multiple distinct classes. The multi-class classifier is trained on labeled 2D data sets to recognize patterns and features that differentiate between multiple predefined classes. This approach enhances the precision of classification tasks by leveraging advanced machine learning algorithms capable of handling multi-class scenarios. The method may involve preprocessing the 2D data to extract relevant features, training the classifier on these features, and applying the trained model to new, unlabeled 2D data sets for classification. The use of a multi-class classifier ensures that the system can accurately assign each input to one of several possible classes, improving decision-making in applications like medical imaging, industrial inspection, or autonomous systems. The technique may also include validation steps to assess the classifier's performance and refine its accuracy over time.
14. The method of claim 13 , wherein a multi-class classifier comprises a support vector machine or a random decision tree.
This invention relates to machine learning classification systems, specifically methods for improving the accuracy and efficiency of multi-class classification tasks. The problem addressed is the challenge of accurately categorizing data into multiple predefined classes, particularly when dealing with complex or high-dimensional datasets where traditional classification methods may struggle with performance or scalability. The invention describes a method for implementing a multi-class classifier that leverages either a support vector machine (SVM) or a random decision tree (RDT) to enhance classification performance. Support vector machines are supervised learning models that identify optimal hyperplanes to separate classes, while random decision trees are ensemble methods that use multiple decision trees to improve robustness and generalization. The method involves training the classifier on labeled data, where the SVM or RDT learns to distinguish between multiple classes by analyzing input features and their relationships. During operation, the trained classifier processes new, unlabeled data to predict the most likely class membership based on learned patterns. The use of SVMs or RDTs in this context provides advantages such as improved accuracy, better handling of high-dimensional data, and reduced overfitting compared to simpler classifiers. The method is particularly useful in applications like image recognition, text categorization, and bioinformatics, where multi-class classification is critical. The invention ensures that the classifier can adapt to different types of data and classification problems, making it versatile for various real-world applications.
15. The method of claim 13 , wherein the multi-class classifier is generated from multiple training sets.
This invention relates to machine learning systems, specifically methods for generating multi-class classifiers from multiple training sets. The core problem addressed is improving classifier accuracy and robustness by leveraging diverse training data sources. The method involves creating a multi-class classifier by training it on multiple distinct training sets. Each training set contains labeled data representing different classes or categories. By combining these sets, the classifier learns to distinguish between multiple classes more effectively than if trained on a single dataset. This approach helps mitigate biases and overfitting that can occur when training on limited or homogeneous data. The training process may involve techniques such as ensemble learning, where multiple models are trained on different subsets of data and their predictions are combined. Alternatively, the classifier may be trained on a merged dataset formed by concatenating the multiple training sets. The method ensures that the classifier can generalize well across different data distributions, improving performance in real-world applications where input data may vary. This technique is particularly useful in domains where data is collected from multiple sources or where class distributions differ significantly between datasets. By incorporating diverse training data, the classifier becomes more adaptable and reliable in practical deployments.
16. The method of claim 1 , wherein the consolidated 3D model is displayed or stored.
A method for generating and utilizing a consolidated 3D model addresses the challenge of integrating multiple 3D data sources into a unified representation for visualization or storage. The method involves acquiring 3D data from various sources, such as sensors, scanners, or databases, and processing this data to correct misalignments, fill gaps, and enhance accuracy. The processed data is then combined into a single 3D model, which may include geometric, textural, or other attribute information. This consolidated model can be displayed on a graphical interface for real-time analysis or stored in a digital format for later use. The method ensures that the final 3D model is coherent, complete, and suitable for applications in fields like engineering, medicine, or virtual reality. By integrating disparate 3D datasets, the method improves efficiency in data handling and enables more accurate modeling and simulation.
17. The method of claim 1 , wherein the correlating locations in the classified 2D data sets to the consolidated 3D data set is based upon a virtual voxel grid for each classified 2D data set.
This invention relates to a method for correlating classified two-dimensional (2D) data sets with a consolidated three-dimensional (3D) data set using a virtual voxel grid. The technology addresses the challenge of accurately aligning and integrating multiple 2D data sets into a unified 3D representation, which is critical in applications such as medical imaging, remote sensing, and industrial inspection where precise spatial correlation is required. The method involves generating a virtual voxel grid for each classified 2D data set. Each voxel grid represents a discrete 3D space where the 2D data points are mapped. The classified 2D data sets are processed to identify features or landmarks that can be matched to corresponding locations in the consolidated 3D data set. The virtual voxel grid facilitates this correlation by providing a structured framework that aligns the 2D data points with the 3D space, ensuring accurate spatial registration. The consolidated 3D data set is constructed by integrating the correlated 2D data sets, leveraging the voxel grid to maintain spatial consistency. This approach improves the accuracy and reliability of the 3D reconstruction, reducing errors that may arise from direct alignment methods. The use of a virtual voxel grid enhances computational efficiency and scalability, making the method suitable for large-scale data processing tasks. The invention is particularly useful in scenarios where high-precision 3D modeling is required from multiple 2D inputs.
18. The method of claim 1 , wherein the classified 2D data sets are further classified as cap.
A system and method for classifying two-dimensional (2D) data sets, such as images or sensor data, to identify and categorize specific features or objects within the data. The method involves processing raw 2D data through an initial classification stage to generate classified 2D data sets, which are then further refined and categorized into more specific sub-classes. In one implementation, the classified 2D data sets are further classified as "cap," indicating a specific type of object or feature detected within the data. This secondary classification may involve additional processing steps, such as pattern recognition, machine learning algorithms, or rule-based filtering, to distinguish between different sub-categories of the initially classified data. The system may be applied in various fields, including industrial inspection, medical imaging, or autonomous navigation, where accurate and detailed classification of 2D data is essential for decision-making or further analysis. The method improves upon existing techniques by providing a more granular and precise classification of detected features, reducing false positives and enhancing the reliability of the system.
19. A quality check module adapted to characterize a specimen and specimen container, comprising: a plurality of cameras arranged around the specimen container and configured to capture multiple images of the specimen container and specimen from multiple viewpoints, each of the plurality of cameras adapted to generate a plurality of 2D images taken at multiple different exposure times and multiple different wavelengths or one or more wavelength ranges; a computer coupled to the plurality of cameras and adapted to process image data from the plurality of 2D images, the computer configured and capable of being operated to: provide classified 2D data sets obtained by processing the plurality of 2D images taken from multiple viewpoints, the classified 2D data sets being classified as: serum or plasma portion, settled blood portion, gel separator (if present), air, tube, and label; correlate locations in the 2D data sets to a consolidated 3D data set; and form a consolidated 3D model based upon the consolidated 3D data set.
This invention relates to a quality check module for characterizing biological specimens and their containers, such as blood collection tubes. The system addresses the need for automated, high-accuracy inspection of specimen containers to ensure proper sample integrity and separation, which is critical for diagnostic testing. The module uses multiple cameras positioned around the container to capture 2D images from various angles, with each camera generating images at different exposure times and wavelengths or wavelength ranges. A computer processes these images to classify different components within the container, including serum or plasma, settled blood, gel separators (if present), air, the tube itself, and any labels. The system correlates spatial data from the 2D images to construct a consolidated 3D data set, which is then used to form a detailed 3D model of the container and its contents. This model enables precise analysis of specimen separation, container integrity, and labeling accuracy, improving quality control in laboratory settings. The multi-spectral and multi-exposure imaging enhances detection of subtle features, while the 3D reconstruction provides comprehensive spatial context for automated quality assessment.
20. A specimen testing apparatus adapted to image a specimen contained within a specimen container, comprising: a track; a carrier on the track configured to contain the specimen container; a plurality of cameras arranged around the track and configured to capture a plurality of 2D images of the specimen container and specimen from multiple viewpoints, each of the plurality of cameras configured to generate a plurality of images at multiple different exposure times and multiple different wavelengths or one or more wavelength ranges; a computer coupled to the plurality of cameras and adapted to process image data from the plurality of 2D images, the computer configured and capable of being operated to: provide classified 2D data sets obtained by processing the plurality of 2D images taken from the multiple viewpoints, the classified 2D data sets being classified as: serum or plasma portion, settled blood portion, gel separator (if present), air, tube, and label; correlate locations in the 2D data sets to a consolidated 3D data set; and form a consolidated 3D model based upon the consolidated 3D data set.
The invention relates to a specimen testing apparatus designed to image and analyze biological specimens, such as blood, contained within specimen containers like test tubes. The apparatus addresses the challenge of accurately identifying and classifying different components within a specimen, such as serum, plasma, settled blood, gel separators, air, and container elements, while also reconstructing a 3D model of the specimen for further analysis. The apparatus includes a track with a carrier that holds the specimen container. Multiple cameras are positioned around the track to capture 2D images of the container and its contents from various angles. These cameras operate at different exposure times and wavelengths or wavelength ranges to enhance image clarity and detail. A computer system processes the captured images, classifying them into distinct data sets based on the identified components (e.g., serum, settled blood, gel separator, air, tube, and label). The system then correlates these 2D data sets to generate a consolidated 3D data set, which is used to form a 3D model of the specimen. This model allows for precise analysis of the specimen's composition and structure, improving diagnostic accuracy and automation in laboratory settings.
Unknown
August 18, 2020
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.